Periodontal disease progression requires the simultaneous presence of high numbers of pathogens, low numbers of compatible or beneficial species, a conductive local environment, and a susceptible host. Effective therapy acts by altering one or more of these factors. Data from an ongoing study were used to examine the biological basis of treatment success or failure. Seventeen subjects showing disease progression were treated by Widman flap surgery at deep sites, scaling at shallow sites, and 1 of 4 randomly-assigned, systemically-administered adjunctive agents including amoxicillin/clavulanate potassium (Au) (n = 3), ibuprofen (n = 3), tetracycline (n = 9), or a placebo (n = 2). Clinical measurements and microbiological samples (enumerated using DNA probes) taken from the mesial aspect of each tooth pre-treatment and 12 months post-treatment were compared and 418 pre- and 418 post-therapy plaque samples were enumerated. Overall, the 4 treatments resulted in pocket depth reduction and "gain" in attachment. After therapy, the percentage of sites colonized by Porphyromonas gingivalis, Prevotella intermedia, Prevotella nigrescens, and Bacteroides forsythus was decreased and counts > 10(6) were less frequent. Large attachment level gains were accompanied by major decreases in these species and were more frequent in subjects receiving antibiotics. A small number of sites in each treatment group became deeper and/or lost attachment. More than half of these sites were detected in 2 subjects who were older (65 vs. 44), had higher serum antibody to Actinobacillus actinomycetemcomitans serotype a (506 vs. 125 ELISA units), A. actinomycetemcomitans serotype b (518 vs. 130), and Campylobacter rectus (39 vs. 18).(ABSTRACT TRUNCATED AT 250 WORDS)